maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/test/CodeGen/AArch64/arm64-inline-asm.ll

235 lines
7.3 KiB
LLVM
Raw Normal View History

[AArch64] Reduce number of callee-save save/restores. Summary: Before this change, callee-save registers would be rounded up to even pairs of GPRs and FPRs. This change eliminates these extra padding load/stores, though it does keep the stack allocation the same size unless both the GPR and FPR sets have an odd size, in which case one full pair stack slot (16 bytes) is saved. This optimization cannot currently be done for MachO targets since they rely on a fast-path .debug_frame equivalent that can only encode callee-save registers as pairs. Reviewers: t.p.northover, rengolin, mcrosier, jmolloy Subscribers: aemerson, rengolin, mcrosier, llvm-commits Differential Revision: http://reviews.llvm.org/D17000 llvm-svn: 260689
2016-02-13 00:31:41 +08:00
; RUN: llc < %s -mtriple=arm64-apple-ios -aarch64-neon-syntax=apple -no-integrated-as -disable-post-ra | FileCheck %s
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
; rdar://9167275
define i32 @t1() nounwind ssp {
entry:
; CHECK-LABEL: t1:
; CHECK: mov {{w[0-9]+}}, 7
%0 = tail call i32 asm "mov ${0:w}, 7", "=r"() nounwind
ret i32 %0
}
define i64 @t2() nounwind ssp {
entry:
; CHECK-LABEL: t2:
; CHECK: mov {{x[0-9]+}}, 7
%0 = tail call i64 asm "mov $0, 7", "=r"() nounwind
ret i64 %0
}
define i64 @t3() nounwind ssp {
entry:
; CHECK-LABEL: t3:
; CHECK: mov {{w[0-9]+}}, 7
%0 = tail call i64 asm "mov ${0:w}, 7", "=r"() nounwind
ret i64 %0
}
; rdar://9281206
define void @t4(i64 %op) nounwind {
entry:
; CHECK-LABEL: t4:
; CHECK: mov x0, {{x[0-9]+}}; svc #0
%0 = tail call i64 asm sideeffect "mov x0, $1; svc #0;", "=r,r,r,~{x0}"(i64 %op, i64 undef) nounwind
ret void
}
; rdar://9394290
define float @t5(float %x) nounwind {
entry:
; CHECK-LABEL: t5:
; CHECK: fadd {{s[0-9]+}}, {{s[0-9]+}}, {{s[0-9]+}}
%0 = tail call float asm "fadd ${0:s}, ${0:s}, ${0:s}", "=w,0"(float %x) nounwind
ret float %0
}
; rdar://9553599
define zeroext i8 @t6(i8* %src) nounwind {
entry:
; CHECK-LABEL: t6:
; CHECK: ldtrb {{w[0-9]+}}, [{{x[0-9]+}}]
%0 = tail call i8 asm "ldtrb ${0:w}, [$1]", "=r,r"(i8* %src) nounwind
ret i8 %0
}
define void @t7(i8* %f, i32 %g) nounwind {
entry:
%f.addr = alloca i8*, align 8
store i8* %f, i8** %f.addr, align 8
; CHECK-LABEL: t7:
; CHECK: str {{w[0-9]+}}, [{{x[0-9]+}}]
call void asm "str ${1:w}, $0", "=*Q,r"(i8** %f.addr, i32 %g) nounwind
ret void
}
; rdar://10258229
; ARM64TargetLowering::getRegForInlineAsmConstraint() should recognize 'v'
; registers.
define void @t8() nounwind ssp {
entry:
; CHECK-LABEL: t8:
; CHECK: stp {{d[0-9]+}}, {{d[0-9]+}}, [sp, #-16]
tail call void asm sideeffect "nop", "~{v8}"() nounwind
ret void
}
define i32 @constraint_I(i32 %i, i32 %j) nounwind {
entry:
; CHECK-LABEL: constraint_I:
%0 = tail call i32 asm sideeffect "add ${0:w}, ${1:w}, $2", "=r,r,I"(i32 %i, i32 16773120) nounwind
; CHECK: add {{w[0-9]+}}, {{w[0-9]+}}, #16773120
%1 = tail call i32 asm sideeffect "add ${0:w}, ${1:w}, $2", "=r,r,I"(i32 %i, i32 4096) nounwind
; CHECK: add {{w[0-9]+}}, {{w[0-9]+}}, #4096
ret i32 %1
}
AArch64: fix conversion of 'J' inline asm constraints. 'J' represents a negative number suitable for an add/sub alias instruction, but while preparing it to become an int64_t we were mangling the sign extension. So "i32 -1" became 0xffffffffLL, for example. Should fix one half of PR20456. llvm-svn: 214052
2014-07-27 15:10:29 +08:00
define i32 @constraint_J(i32 %i, i32 %j, i64 %k) nounwind {
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
entry:
; CHECK-LABEL: constraint_J:
%0 = tail call i32 asm sideeffect "sub ${0:w}, ${1:w}, $2", "=r,r,J"(i32 %i, i32 -16773120) nounwind
AArch64: fix conversion of 'J' inline asm constraints. 'J' represents a negative number suitable for an add/sub alias instruction, but while preparing it to become an int64_t we were mangling the sign extension. So "i32 -1" became 0xffffffffLL, for example. Should fix one half of PR20456. llvm-svn: 214052
2014-07-27 15:10:29 +08:00
; CHECK: sub {{w[0-9]+}}, {{w[0-9]+}}, #-16773120
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
%1 = tail call i32 asm sideeffect "sub ${0:w}, ${1:w}, $2", "=r,r,J"(i32 %i, i32 -1) nounwind
AArch64: fix conversion of 'J' inline asm constraints. 'J' represents a negative number suitable for an add/sub alias instruction, but while preparing it to become an int64_t we were mangling the sign extension. So "i32 -1" became 0xffffffffLL, for example. Should fix one half of PR20456. llvm-svn: 214052
2014-07-27 15:10:29 +08:00
; CHECK: sub {{w[0-9]+}}, {{w[0-9]+}}, #-1
%2 = tail call i64 asm sideeffect "sub ${0:x}, ${1:x}, $2", "=r,r,J"(i64 %k, i32 -1) nounwind
; CHECK: sub {{x[0-9]+}}, {{x[0-9]+}}, #-1
%3 = tail call i64 asm sideeffect "sub ${0:x}, ${1:x}, $2", "=r,r,J"(i64 %k, i64 -1) nounwind
; CHECK: sub {{x[0-9]+}}, {{x[0-9]+}}, #-1
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
ret i32 %1
}
define i32 @constraint_KL(i32 %i, i32 %j) nounwind {
entry:
; CHECK-LABEL: constraint_KL:
%0 = tail call i32 asm sideeffect "eor ${0:w}, ${1:w}, $2", "=r,r,K"(i32 %i, i32 255) nounwind
; CHECK: eor {{w[0-9]+}}, {{w[0-9]+}}, #255
%1 = tail call i32 asm sideeffect "eor ${0:w}, ${1:w}, $2", "=r,r,L"(i32 %i, i64 16711680) nounwind
; CHECK: eor {{w[0-9]+}}, {{w[0-9]+}}, #16711680
ret i32 %1
}
define i32 @constraint_MN(i32 %i, i32 %j) nounwind {
entry:
; CHECK-LABEL: constraint_MN:
%0 = tail call i32 asm sideeffect "movk ${0:w}, $1", "=r,M"(i32 65535) nounwind
; CHECK: movk {{w[0-9]+}}, #65535
%1 = tail call i32 asm sideeffect "movz ${0:w}, $1", "=r,N"(i64 0) nounwind
; CHECK: movz {{w[0-9]+}}, #0
ret i32 %1
}
define void @t9() nounwind {
entry:
; CHECK-LABEL: t9:
%data = alloca <2 x double>, align 16
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%0 = load <2 x double>, <2 x double>* %data, align 16
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
call void asm sideeffect "mov.2d v4, $0\0A", "w,~{v4}"(<2 x double> %0) nounwind
; CHECK: mov.2d v4, {{v[0-9]+}}
ret void
}
define void @t10() nounwind {
entry:
; CHECK-LABEL: t10:
%data = alloca <2 x float>, align 8
%a = alloca [2 x float], align 4
[opaque pointer type] Add textual IR support for explicit type parameter to getelementptr instruction One of several parallel first steps to remove the target type of pointers, replacing them with a single opaque pointer type. This adds an explicit type parameter to the gep instruction so that when the first parameter becomes an opaque pointer type, the type to gep through is still available to the instructions. * This doesn't modify gep operators, only instructions (operators will be handled separately) * Textual IR changes only. Bitcode (including upgrade) and changing the in-memory representation will be in separate changes. * geps of vectors are transformed as: getelementptr <4 x float*> %x, ... ->getelementptr float, <4 x float*> %x, ... Then, once the opaque pointer type is introduced, this will ultimately look like: getelementptr float, <4 x ptr> %x with the unambiguous interpretation that it is a vector of pointers to float. * address spaces remain on the pointer, not the type: getelementptr float addrspace(1)* %x ->getelementptr float, float addrspace(1)* %x Then, eventually: getelementptr float, ptr addrspace(1) %x Importantly, the massive amount of test case churn has been automated by same crappy python code. I had to manually update a few test cases that wouldn't fit the script's model (r228970,r229196,r229197,r229198). The python script just massages stdin and writes the result to stdout, I then wrapped that in a shell script to handle replacing files, then using the usual find+xargs to migrate all the files. update.py: import fileinput import sys import re ibrep = re.compile(r"(^.*?[^%\w]getelementptr inbounds )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))") normrep = re.compile( r"(^.*?[^%\w]getelementptr )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))") def conv(match, line): if not match: return line line = match.groups()[0] if len(match.groups()[5]) == 0: line += match.groups()[2] line += match.groups()[3] line += ", " line += match.groups()[1] line += "\n" return line for line in sys.stdin: if line.find("getelementptr ") == line.find("getelementptr inbounds"): if line.find("getelementptr inbounds") != line.find("getelementptr inbounds ("): line = conv(re.match(ibrep, line), line) elif line.find("getelementptr ") != line.find("getelementptr ("): line = conv(re.match(normrep, line), line) sys.stdout.write(line) apply.sh: for name in "$@" do python3 `dirname "$0"`/update.py < "$name" > "$name.tmp" && mv "$name.tmp" "$name" rm -f "$name.tmp" done The actual commands: From llvm/src: find test/ -name *.ll | xargs ./apply.sh From llvm/src/tools/clang: find test/ -name *.mm -o -name *.m -o -name *.cpp -o -name *.c | xargs -I '{}' ../../apply.sh "{}" From llvm/src/tools/polly: find test/ -name *.ll | xargs ./apply.sh After that, check-all (with llvm, clang, clang-tools-extra, lld, compiler-rt, and polly all checked out). The extra 'rm' in the apply.sh script is due to a few files in clang's test suite using interesting unicode stuff that my python script was throwing exceptions on. None of those files needed to be migrated, so it seemed sufficient to ignore those cases. Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7636 llvm-svn: 230786
2015-02-28 03:29:02 +08:00
%arraydecay = getelementptr inbounds [2 x float], [2 x float]* %a, i32 0, i32 0
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%0 = load <2 x float>, <2 x float>* %data, align 8
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
call void asm sideeffect "ldr ${1:q}, [$0]\0A", "r,w"(float* %arraydecay, <2 x float> %0) nounwind
; CHECK: ldr {{q[0-9]+}}, [{{x[0-9]+}}]
call void asm sideeffect "ldr ${1:d}, [$0]\0A", "r,w"(float* %arraydecay, <2 x float> %0) nounwind
; CHECK: ldr {{d[0-9]+}}, [{{x[0-9]+}}]
call void asm sideeffect "ldr ${1:s}, [$0]\0A", "r,w"(float* %arraydecay, <2 x float> %0) nounwind
; CHECK: ldr {{s[0-9]+}}, [{{x[0-9]+}}]
call void asm sideeffect "ldr ${1:h}, [$0]\0A", "r,w"(float* %arraydecay, <2 x float> %0) nounwind
; CHECK: ldr {{h[0-9]+}}, [{{x[0-9]+}}]
call void asm sideeffect "ldr ${1:b}, [$0]\0A", "r,w"(float* %arraydecay, <2 x float> %0) nounwind
; CHECK: ldr {{b[0-9]+}}, [{{x[0-9]+}}]
ret void
}
define void @t11() nounwind {
entry:
; CHECK-LABEL: t11:
%a = alloca i32, align 4
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%0 = load i32, i32* %a, align 4
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
call void asm sideeffect "mov ${1:x}, ${0:x}\0A", "r,i"(i32 %0, i32 0) nounwind
; CHECK: mov xzr, {{x[0-9]+}}
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%1 = load i32, i32* %a, align 4
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
call void asm sideeffect "mov ${1:w}, ${0:w}\0A", "r,i"(i32 %1, i32 0) nounwind
; CHECK: mov wzr, {{w[0-9]+}}
ret void
}
define void @t12() nounwind {
entry:
; CHECK-LABEL: t12:
%data = alloca <4 x float>, align 16
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%0 = load <4 x float>, <4 x float>* %data, align 16
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
call void asm sideeffect "mov.2d v4, $0\0A", "x,~{v4}"(<4 x float> %0) nounwind
Add a bunch of CHECK missing colons in tests. NFC. Some wouldn't pass; fixed most, the rest will be fixed separately. llvm-svn: 232239
2015-03-14 09:43:57 +08:00
; CHECK: mov.2d v4, {{v([0-9])|(1[0-5])}}
ARM64: initial backend import This adds a second implementation of the AArch64 architecture to LLVM, accessible in parallel via the "arm64" triple. The plan over the coming weeks & months is to merge the two into a single backend, during which time thorough code review should naturally occur. Everything will be easier with the target in-tree though, hence this commit. llvm-svn: 205090
2014-03-29 18:18:08 +08:00
ret void
}
define void @t13() nounwind {
entry:
; CHECK-LABEL: t13:
tail call void asm sideeffect "mov x4, $0\0A", "N"(i64 1311673391471656960) nounwind
; CHECK: mov x4, #1311673391471656960
tail call void asm sideeffect "mov x4, $0\0A", "N"(i64 -4662) nounwind
; CHECK: mov x4, #-4662
tail call void asm sideeffect "mov x4, $0\0A", "N"(i64 4660) nounwind
; CHECK: mov x4, #4660
call void asm sideeffect "mov x4, $0\0A", "N"(i64 -71777214294589696) nounwind
; CHECK: mov x4, #-71777214294589696
ret void
}
define void @t14() nounwind {
entry:
; CHECK-LABEL: t14:
tail call void asm sideeffect "mov w4, $0\0A", "M"(i32 305397760) nounwind
; CHECK: mov w4, #305397760
tail call void asm sideeffect "mov w4, $0\0A", "M"(i32 -4662) nounwind
; CHECK: mov w4, #4294962634
tail call void asm sideeffect "mov w4, $0\0A", "M"(i32 4660) nounwind
; CHECK: mov w4, #4660
call void asm sideeffect "mov w4, $0\0A", "M"(i32 -16711936) nounwind
; CHECK: mov w4, #4278255360
ret void
}
define void @t15() nounwind {
entry:
%0 = tail call double asm sideeffect "fmov $0, d8", "=r"() nounwind
; CHECK: fmov {{x[0-9]+}}, d8
ret void
}
; rdar://problem/14285178
define void @test_zero_reg(i32* %addr) {
; CHECK-LABEL: test_zero_reg:
tail call void asm sideeffect "USE($0)", "z"(i32 0) nounwind
; CHECK: USE(xzr)
tail call void asm sideeffect "USE(${0:w})", "zr"(i32 0)
; CHECK: USE(wzr)
tail call void asm sideeffect "USE(${0:w})", "zr"(i32 1)
; CHECK: orr [[VAL1:w[0-9]+]], wzr, #0x1
; CHECK: USE([[VAL1]])
tail call void asm sideeffect "USE($0), USE($1)", "z,z"(i32 0, i32 0) nounwind
; CHECK: USE(xzr), USE(xzr)
tail call void asm sideeffect "USE($0), USE(${1:w})", "z,z"(i32 0, i32 0) nounwind
; CHECK: USE(xzr), USE(wzr)
ret void
}